Team Project

Pre-diagnostic Pancreatic Cancer Set Aside Project

PaCa Pre-Dx Bake Off 3
Feng, ZidingFred Hutchinson Cancer Research Center
Van Andel: CA 19-9, sTRA, sTRA-glycan. MDACC: CA19-9,LRG1,TIMP1. MDAC: Metabolomic analyses for referenced targets of interest (acetylspermidine, diacetylspermine, an indole-derivative, and two lysophosphatidylcholines (LPCs). UNMC: MUC4, MUC5AC, CA19. UPMC: CA 19-9, CEA, NSE, bHCG, CEACAM1, PRL, THSP, CA125, TIMP1, TIMP3, IGFBP1,3, OPN, CATHD, OPG, CRP, ANGPTL3,4, SAA, VCAM1, APOA1. FHCRC: Triplicate features of each antibody or protein are printed onto Schott Nexterion H slides using our Genetix Q-Array 2 microarray platform.
No design specified.
G.I. and Other Associated Cancers Research Group

For the early detection of pancreatic cancer, screening among the general population is not feasible due to low prevalence of the disease in general population. An alternative strategy is surveillance for incipient pancreatic cancer among a population with elevated risk. An elevated-risk condition that has gained attention in recent years is sudden-onset type-2 diabetes (1). In that group, the prevalence of pancreatic cancer may be as high as 0.8% (2). At such a prevalence, a biomarker with 96% specificity and 65% sensitivity would have a positive predictive value (PPV) of 11.6% and negative predictive value (NPV) of 99.7%, which could be acceptable in a cost-benefit analysis. The clinical application for using the biomarker among high-risk population (NoD) is to recommend imaging workup if the biomarker tests positive. Subjects with positive imaging result will then move forward in evaluation and treatment as each case requires. The level of PPV proposed here puts the yield of cancer in the test-positive group compatible with the yield of colorectal cancer screening by fecal occult-blood tests for people aged > 50 (PPV=2.5-5.0%, sensitivity=69-79%), or with the yield of lung cancer LDCT screening for heavy smokers (PPV=3.6%).

Aim 1. Assemble and distribute samples from pre-diagnostic cohorts Aim 2. Test the individual and combination markers that were suggested by the previous team projects in pre-diagnostic cohorts
Dr. Huang at the FHCRC will perform the statistical analyses. The goal will be to determine performance of each biomarker or biomarker panel and whether each panel passes the minimumperformance requirement. Receiver operating characteristic curve will be estimated for each individual marker and CA199, with 95% confidence interval for each point on the ROC curve estimated using 1000 bootstrap resamples. If the lower bound of the 95% confidence interval of the tested biomarker exceeds the minimum performance criterion (40% sensitivity at 95% specificity), the biomarker will have passed the performance criterion. For comparison with CA19-9, the panel will have passed the performance criterion if the lower bound of 95% confidence interval of the difference in performance between the panel and CA19-9 exceeds zero. For biomarkers not meeting the target performance, we will perform secondary analyses to determine the cause of the failure and strategies for further development (see alternative strategies). Biomarker panels cross labs will also be developed using various algorithms, including logistic regression (for its simplicity and oftentimes relatively robust performance in classification), classification tree, and support vector machine. Best panel will be selected based on cross-validated performance. Panels will be developed for specimens collected within different periods prior to disease diagnosis (<1yr, <2yr, <3yr, <4yr, <5yr).

No datasets are currently associated with this protocol.

Version 5.1.0